摘要
汽车座椅舒适性的主观评价和客观评价之间的关系存在复杂性和高非线性的特点。传统的BP神经网络(反向传播神经网络)对初值敏感且容易陷入局部最优解,因此无法建立精确的座椅舒适度预测模型。针对该问题,提出利用人工蜂群算法优化后的BP神经网络(ABC-BP)来预测座椅的舒适性的方法。通过开展体压试验得到了176组压力分布样本数据,其中89%的数据作为模型的训练部分,11%的数据作为模型验证,将预测结果与真实值相比较,ABC-BP预测模型的均方误差MSE为0.0019,确定性系数R^2为0.946,比传统BP神经网络预测模型得到的MSE降低了84.68%,R^2提高了42.5%。结果表明,利用人工蜂群算法优化后的BP神经网络所建立的汽车座椅舒适性预测模型稳定性更强、预测效果更加精准。
The relationship between subjective and objective evaluation of automobile seat comfort has the characteristics of complexity and high nonlinearity.The traditional BP neural networks(back propagation neural networks)are sensitive to weighted initial values and easily converge to local minimal value.Therefore,it is difficult to establish an accurate seat comfort prediction model.To solve the problem,this paper proposes a method for predicting automobile seat comfort by using the BP neural networks optimized by artificial bee colony algorithm.176 groups of pressure distribution sample data were obtained through automobile body pressure test,89%of which were used to train the model;11%were used as model validation.Compared with the results predicted with real values,the results predicted with our method reach 0.0019 in mean square error and 0.946 in determination coefficient;the mean square error is 84.68%lower than that obtained with the BP neural network algorithm and the R^2(determination coefficient)is 42.5%higher than that obtained with the BP neural network algorithm.The results show that the prediction model of automobile seat comfort based on the BP neural networks optimized by the artificial bee colony algorithm is more stable and accurate.
作者
龙江
郭鹏程
陈梓铭
陈翼华
李落星
Long Jiang;Guo Pengcheng;Chen Ziming;Chen Yihua;Li Luoxing(State Key Laboratory of Design and Manufacture for Vehicle Body,Hunan University,Changsha 410082,China;Oushang Institute of Chang'an Automobile,Chongqing 400023,China)
出处
《机械科学与技术》
CSCD
北大核心
2020年第2期273-281,共9页
Mechanical Science and Technology for Aerospace Engineering
基金
国家重点研发计划项目(2016YFB0101700)资助.
关键词
人工蜂群
神经网络
压力分布
汽车座椅舒适性
artificial bee colony algorithm
neural networks
pressure distribution
automobile seat comfort
